Proposal of deployment reconfiguration for environment adaptation

TL;DR

This paper introduces a method for dynamic relocation of environment-adaptive software during operation, optimizing placement considering multiple users to enhance overall satisfaction using linear programming.

Contribution

It proposes a novel relocation approach during runtime for environment-adaptive software, considering multi-user placement for improved performance.

Findings

Effective rearrangement confirmed through simulation

Improves overall user satisfaction

Utilizes linear programming for placement optimization

Abstract

To use heterogeneous hardware, programmers needed sufficient technical skills such as OpenMP, CUDA, and OpenCL. Therefore, I have proposed environment-adaptive software that enables high-performance operation by automatically converting and configuring the code once written, and have been working on automatic conversion and proper placement. However, until now, where to initially place the converted application has been considered, but the overall optimal placement has not been considered in consideration of the placement status of other users. In this paper, as a new element of environment-adaptive software, I study the relocation during operation, which improves the overall user satisfaction by considering the placement of other users, using a linear programming method. It was confirmed that it can be properly rearranged through simulation experiments.